-
Cloning and Immunolocalization
Andreas Jeromin
Department of Physiology
A thesis submitted in conformity with the requirements for the Degree of the Master of Science in the
University of Toronto
O Andreas Jeromin 1998
-
National Library Bibliothèque nationale du Canada
Acquisitions and Acquisitions et Bibliog raphic Services seMces bibliographiques
395 Wellington Street 395. nie Wellington OtbwaON K1AûN4 OttawaON KiAON4 canada Canada
The author has granted a non- L'auteur a accordé une licence non exclusive licence ailowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thests in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de
reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fkom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.
-
A bstract
Frequenin in Crustaceans : Cloning and Immunolocalization, Master of Science, 1998,
Andreas Jeromin, Department of Physiology, University of Toronto.
Crustaceans provide many advantages in the snidy of synaptic transmission and plasticity.
Neuromuscular preparations have been well characterized in terms of long-term
facilitation, synaptic depression and long-term adaptation. In addition, crustacean motor
neurons are excellent models to study the differentiation into 'phasic' and 'tonic' motor
neurons. These two types of rnotor neurons differ substantially in their transmitter release
properties. Although the ultrastructural and biochemical differences between these neurons
have been well studied, very Iittle is known about possible rnolecular factors underlying
this differentiation.
The neuronal calcium-binding protein frequenin has been shown to enhance synaptic
eficacy at neuromuscular junction in Drosophila and Xenopus and could serve as such a
molecular factor. Standard molecular biology techniques were used to clone frequenin from
lobster and crayfkh. The lobster and crayfish frequenins were found to be closely related to
originally identified Drosophila frequenin. In addition, immunolocalization of frequenin at
the crayfish neuromuscular junction revealed that frequenin is heavily expressed in
'phasic'and less strongly in 'tonic' motor neurons.
-
Acknowledgements
1 would like to thank my supervisor, Dr. H.L. Atwood, for the opportunity to punue this
project and for
immunoc ytochemical
his support and guidanceThe craflsh and Drosophila
experimentation, confocal irnaging, and analysis were done as a
- shared collaborative effort with Mr. Ai Shayan. 1 thank Mrs. Marianne Hegstr m-
Wojtowicz for technical assistance with imrnunocytochernistry. Dr. Roder is thanked for
providing access to his molecular biology facilities at the Samuel Lunenfeld Research
Institute (SLRI), Mount Sinai Hospital, Toronto, and for his support. Wanda helped me to
get over many of the nitty gritties of the daily lab routine.
I also thank the members of my supervisory cornmittee, Drs. Charlton and Roder for
sharing their insight, and Dr. J.F. MacDonald for chairing rny thesis review.
Drs. H. Bellen, E. Buchner, R. Kelly, OPongs, J. Roos and K. Zinsmaier are
thanked for providing invaluable aliquots of the antibodies their labs have raised. Many
thanks to Dr. G. Lnenicka who kindly provided an aliquot of the crayfish nerve-cord library
and to Dr. W.-D. Krenz who provided an aliquot of the Iobster nerve-cord library. Support
from the Medicai Research Council of Canada (research grant to H.L. Atwood and a
studentship to A. Jeromin) is gratefully acknowledged. Finally, very special thanks to my
family for their endless support and encouragement.
-
Table of contents
Abstract
Acknowledgements
Table of Contents
List of Figures
List of Tables
List of Abbreviations
INTRODUCTION
Synaptic transmission at the neuromuscular junction
Crustacean synaptic differentation
Frequenin : a neuronal calcium-binding protein
Objective of the present study
MATERIALS AND METHODS
Chernicals
Animals
Cloning of crayfish frequenin
a.) PCR amplifcation of an intemal DNA fragment of frequenin
-
b.) Screening of the crayfkh nerve cord cDNA library
L) PLating of the library and transfer to membranes
II.) Random-primed labeling of the internai PCR
fragment
Immunocytochernistry
a.) Animds
III.) High-stnngency hybridization in aqueous solution 24
26
26
L ) Crayfkh 26
Il.) Drosophila melanogaster 26
Preparations 26
L) Crayfkh muscle dissections 26
II.) Preparation of Drosophila 3rd instar larvai fillets 27
Solutions 27
1.) Crayfish saline 27
II.) Solution used for Drosophila larval dissections 28
Anti bodies 28
-
e.) Fixation and staining
1.) Prefixation of crayfi~sh muscle preparations
IL) Fixation and staining
III.) Irnaging
29
29
29
3 1
RESULTS 32
a.) Cloning of crayfi~sh frequenin 32
b.) Immunolocalization of synapsin, dynamin and frequenin-like molecules at
crayfish and Drosophila neuromuscular junc tions 60
c.) Immunolocalization at crayfish neuromuscular junctions 6 1
1.) Synapsin 6 1
II.) Dynarnin 64
III.) Frequenin 64
d.) Immunlocaiization of synapsin, dynarnin and frequenin at the Drosophila
larval neuromuscular junction
1.) Synapsin
II.) Dynarnin
III.) Frequenin
-
DISCUSSION 75
a) Cloning of crayfish frequenin 75
1.) General molecular features 75
II.) Conservation of EF-hands 76
IIL) N-myristoylation motif 76
TV.) Functional implications derived from the prirnary
sequence 77
V.) Future work 78
b.) Immunoiocalization of frequenin at the NMJ 80
1.) Patterns of localization 80
IL) Functional implications : physiology 82
III.) Functional implications : morphology 83
c.) Possible cellular mechanisms 85
d.) General conclusions 89
SUMMARY 90
REFERENCES 92
APPENDlX 102
vii
-
List of Figures
Figure 1. Schematic diagram of the procedure used for cloning of crayfish and lobster
frequenin 19
Figure 2. Primers used for PCR amplification 23
Figure 3. Gel electrophoresis of the amplified PCR products from the crayfish nerve-cord
cDNA library 35
Figure 4. Gel electrophoresis of the amplified PCR products from the lobster nerve-cord
cDNA library 37
Figure 5. Analysis of the insert sire of the crafish and lobster frequenin cDNA clones
identified from the high-stringency screen 41
Figure 6. Analysis of the possible open reading frames of the crayfish frequenin cDNA 43
Figure 7. Analysis of the possible open reading frames of the lobster frequenin cDNA 46
Figure 8. Protein sequence of crayfish frequenin 5 1
Figure 9. Protein sequence of lobster frequenin 53
Figure 10. Pair-wise alignment of the crayfi~sh and Drosophila frequenin protein
sequences 55
Figure 11. Pair-wise alignment of the lobster and Drosophila frequenin protein
sequences 57
Figure 12. Pair-wise alignment of the crayfkh and lobster frequenin protein sequences 59
Figure 13. hmunolocaiization of synapsin, dynarnin and frequenin at the neuromuscular
junction of the crayfi~sh leg extensor muscle 68
-
Figure 14. Immunolocalization of synapsin, dynamin and frequenin at the neurornuscular
junction of the crayfkh fast extensor muscles of the abdomen 70
Figure 15. Irnmunolocalization of synapsin, dynamin and frequenin at the NMJ of the
crayfkh slow abdominal flexor muscles
Figure 16. Immunolocalization of synapsin, dynamin and frequenin at the mature
Drosophila larval neuromuscular junc tion
Figure 17. Predicted mode1 structure of crayfish frequenin
-
List of TabIes
Table 1 . Antibodies tested for cross-reactivity at crayti~sh neuromuscular junctions 63
-
List of Abbreviations
BSA
dATP
dCTP
dGTP
d m
EPSP
LTA
NGS
NMJ
PFA
PCR
SSC buffer
Bovine serum alburnin
Deoxyadenosine triphosphate
Deoxycytidine triphosphate
Deoxyguanosine triphosphate
Deoxythymidine triphosphate
Excitatory postsynaptic potential
Long-term adaptation
Normal goat serum
Neuromuscular junction
Paraformaldehyde
Polyrnerase chain reaction
Sodium chloride/ sodium citrate
-
Introduction
Synaptic transmission is one of the basic processes of neuronal communication. This
process is a dynamic activity-dependent phenornenon which is influenced by a variety of
factors , including the development and differentiation of the presynaptic terminal, and the
overall level of nerve impulse activity carried by the neuron (Atwood and Cooper, 1996).
At the neuromuscular junction in crustaceans, synaptic communication between the
presynaptic motoneuron and the postsynaptic muscle is highly differentiated to accomplish
specific physiological functions. Past work has defined many features of the physiology of
these neurons, but relatively little is known about the molecular differences responsible for
their physiological differentiation. The general goal of the present study was to further
define the molecular basis for this synaptic differentiation.
Crustacean motor neurons have provided useful experimental models for
investigating synaptic transmission and synaptic plasticity (Atwood and Wojtowicz, 1986;
Atwood, Cooper, and Wojtowicz, 1994; and Atwood and Cooper, 1995 ). Individual motor
neurons can be identified, are large and easiiy accessible for experimental manipulation,
and show many of the physiological properties observed in other species such as short-tem
-
and long-tem facilitation, and depression. In crustaceans, two types of motor neuron have
been identified that differ fundamentally in their physiological activity and transmitter-
releasing properties. These motor neurons are generally terrned 'phasic' and 'tonic'
(Kennedy and Takeda, 1965 a and b). For a single nerve impulse the 'phasic' motor neuron
can release up to 1000-fold the arnount of transmitter released by the tonic motor neuron
ont0 a single muscle fiber (Atwood et al. 199 1).
In searching for molecular differences which might account for neuronal
differentiation in crustaceans, Atwood et al. (1995~) found that a frequenin-iike molecule
is differentially expressed in crayfish rnotoneuron terminais. This neuronal calcium binding
protein has been implicated as a molecule that modulates synaptic efficacy (Pongs et al.,
1993). Overexpression of frequenin in Drosophila led to increased transmitter release
under conditions of repetitive stimulation (Pongs et al., 1993 ; Rivosecchi et al. 1994).
Olafsson et al. (1995) reported that infusion of Xenopus frequenin into Xenopus motor
neurons increased spontaneous quantal transmission and evoked transmitter release. These
observations suggest that frequenin has a role in regulating the strength of synaptic
transmission. Its differential expression in crustacean motoneurons raises the interesting
possiblity that some of the observed physiological differentiation might be due to this
molecular difference.
-
In the following, 1 will review the general feanires of synaptic transmission at
crustacean neuromuscular junctions and the known properties of frequenin, relating these
to the physiological properties of the 'phasic' and 'tonic' motor neuron, and develop the
hypothesis that frequenin plays a role in the differences in transrnitter release of these two
types of motor neuron.
Synaptic transmission at the neuromuscular junction :
The release of the neurotransrnitter glutamate from the presynaptic terminal
and its subsequent binding to its postsynaptic receptors on the muscle surface is the basic
process of synaptic communication between the motor neuron and the muscle in
crustaceans and other arthropods. The arrangement of a prespaptic release site separated
from the opposing speciaiized postsynaptic membrane by a smdl cleft is defined as a
(chernical) synapse. Freeze-fracture photomicrographs of the presynaptic extracellular
surface of crustacean synapses (Govind et al, 1994) show clusters of large
intramembranous particles, representing putative ca2+ and ~a"-dependent Kf channels (
e.g. Cooper et ai. ,1996), in close association with vesicle fusion points mediating the
release of neurotransrnitter (Heuser & Reese, 1979). Where synaptic vesicles cluster near
the presynaptic thickening, transmission electron micrographs show a localized
accumulation of electron-dense material, the dense body or presynaptic dense bar. Such
structures are comrnonly seen at synapses of crustacean neurornuscular junctions (Jahromi
and Atwood 1974; Walrond et aI. 1993).
-
Molecular biological and biochemical efforts have revealed that vesicular
exocytosis is mediated by a regulated network of protein-protein interactions (Scheller,
1995 ; Sudhof, 1995). A picture is emerging of the molecular machinery by which vesicles
are docked at release sites and membrane fusion is triggered in response to calcium influx.
It has become clear that synaptic vesicle docking and fusion cm be viewed as a special
case of membrane trafficking (for review, see Bennett and Scheller (1993). Selective
targeting of a particular vesicle to a particular membrane destination is thought to be
mediated by specific combinations of vesicle and target membrane proteins (called SNAP
receptors or SNAREs).
According to the current rnodel, spaptic vesicles dock and then proceed
through a partial hsion reaction, priming, to make hem competent for the final
triggered step (Sudhof, 1995). During priming, a complex called the core complex is
assembled from the 3 synaptic proteins, two from the plasma membrane (syntaxin and
SNAP-25) and one from synaptic vesicles (snaptobrevid VAMP). The core complex forms
the anchor for a cascade of protein-protein interactions required for exocytosis to occur.
Once the trimeric core complex has forrned, it serves as a receptor for SNAP and the N-
ethylmaleimide-sensitive factor (NSF). NSF is an ATPase, while the SNAPs are soluble
NSF-attachment proteins. Assembly of the SNAPs into the core complex creates a high-
aff3nity binding site for NSF which then disrupts the complex under ATP hydrolysis. Thus,
an ordered sequence of protein-protein-interactions leads to the assembly of a multimeric
complex which is then disrupted by NSF.
-
To trigger the final reaction of the fusion reaction, a ~ a " sensor is required
at the site of exocytosis (see below). Synaptotagmin which occurs in various tissue-
specific isoforms and has cazf-binding domains has been identified as such a sensor (for
review, see Sudhof (1995). It binds ca2+ cooperatively and undegoes a ca2+-dependent
conformational change which completes the fusion reaction.
Two proteins that are not an integrai part of the SNAREs underlying the
(final) docking and fusion steps are the synapsins and dynamin. Synapsin has been
identified as a molecule important for vesicle mobilization (Pieribone et al. 1996) whereas
dynamin, a GTPase, has been shown to be essentiai for the recycling of fused vesicles
(endocytosis) (Estes et al. 1996).
Exocytosis of neurotransmitter occurs in response to a presynaptic action
potential which leads to a transient increase in the calcium concentration within the nerve
terminal through the opening of voltage-gated calcium channels and their interaction with
synaptic vesicle proteins probably including spaptotagrnin and possibly also syntaxin and
SNAP-25. At al1 synapses, including the crustacean neuromuscular junction, the release of
neurotransrnitter occurs in discrete multimolecular packages, i.e. is quanta1 in nature
(Dudel & Kuffler, 196 1 ). The morphological counterpart to the quantum is thought to be
the single synaptic vesicle (Atwood & Wojtowicz, 1986). Ai the crayfkh neuromuscular
junction, it is believed that there are approximately 6000 molecules of glutamate per
vesicle (Atwood & Wojtowicz, 1986).
Depolarkation-induced ~ a " entry through voltage-gated calcium channels
of the presynaptic terminal has been shown to be a factor for the induction of
-
neurotransmitter release. Katz and Miledi (1967 and 1968) demonstrated that aitering the
concentration of extemal ~ a ' + affected the arnount of transmitter released by nerve
terminal depolarkation. Direct evidence for a positive relationship between Ca2+ influx and
the magnitude of neurotransmitter secretion has been shown in studies of the giant synapse
in squid (for review, see Augustine et al. 199 1).
Current anatornical, physiological and theoretical findings indicate that
calcium channels are clustered close to specific sites, the active zones. at which vesicles
release their contents and have reinforced the concept that ~ a ' + entenng thorough voltage-
gated calcium channels exists iransiently in a locaiized spatial domain, the so-called ~ a ' +
microdomain (for review, see Stanley et al. 1997). It is becoming clear that the spatial-
temporal dynamics of these ca2+ microdomains determine the arnount of transmitter
released. The interaction between these Ca" microdomains by diffusion is affected by
mobile and immobile Ca2+ buffer present and could present a mechanism by which more
and more active zones becom5 recruited during repetitive firing of nerve impulses (Cooper
et al. 1996)
As a calcium-binding protein. frequenin could alter the dynarnics of ca2'
in these microdomains or its entry through voltage-gated channels. Although frequenin is
unlikely to act as mere ca2+ buffer for reasons outlined below, it will be very informative
-
to study the dynamics of ~ a " in the presynaptic terminal in which the expression of
frequenin has k e n altered.
Cmstacean synaptic differentiation :
Two broad classes of motor neurons have been shown to exist in the crustaceans. These
have k e n designated as 'phasic' and 'tonic' types (Kennedy and Takeda 1965 a,b). Since
these motor neurons are physiologicaliy and rnorphologically distinct from one another.
they serve as suitable models for studying factors underlying synaptic differentiation
(Cooper and Atwood, 1995).
The differences between these motor neurons are related to the physiological
demands of the muscle fibers that they innervate. 'Phasic' neurons innervating the fast-
acting muscles of the abdomen in crustaceans are generally silent and are recruited for
rapid swimming and escape responses. The terminais of the 'phasic' rnotor neurons
produce large excitatory post-synaptic potentials (EPSPs) which Iead to evoked action
potentials and hence contraction of the muscle fibers. Repetitive stimulation of the
'phasic' neurons leads to a rapid depression of EPSPs. These physiological effects are
attributable to an initially large quantal output of transmitter and its subsequent decline (for
review, see Atwood and Cooper, 1996).
-
'Tonic' neurons innervating the slow-acting postural muscles of the abdomen are
usually active much of the time. Repetitive stimulation of the 'tonic' motor neurons at
high frequency leads to facilitation rather than depression. The muscle responsible for the
opening of the claw (hence called the claw opener muscle) receives its only excitatory
input from a single axon with 'tonic' properties while the claw closer muscle and
carpopodite extensor muscle of the wdking limb is conjointly innervated by both 'phasic'
and 'tonic' axons. in the claw cioser muscle , the rapidly contracting fibers receive strong
input from the 'phasic' axon while the slowly contracting fibers receive strong 'tonic' input
(Atwood, 1982).
It is known from previous studies that there is an overall rnorphologicai difference in
the nerve terminal morphology of these two types of axons. 'Phasic' axon terminals are
relatively filiform and non-varicose, while 'tonic' axon terminais are larger in diameter and
distinctly varicose (Lnenicka et al., 1986, 199 1; Bradacs et al., 1997). At the
ultrastrucniral level, the terminals of the 'phasic' rnotor neurons have uniformly-distributed
synapses, whereas the synapses in the 'tonic' nerve endings are less widely dispersed and
more confined in their localization to varicosities. Recordings from single boutons with
focal extracellular electrodes have demonstrated differences in the quantal content of
'phasic' and 'tonic' motor neurons with the 'phasic' terminais releasing 50-2000 times
-
more transrnitter per nerve impulse than their 'tonic' counterparts. While on average there
is more synaptic surface area per unit length of the 'tonic' terminal (King et al. 1996)- the
release of quanta per single action potential is much greater in the filiform 'phasic' nerve
endings (Cooper et al., L995b; Bradacs et al., 1996). These findings suggest that the
observed outstanding ultrastructurai differences aione cannot account for the much greater
transmitter output of 'phasic' nerve terminais when compared to their 'tonic' counterparts.
Differences in electrical and biochemical properties rather than gross ultrastructural
differences are more likely to play important roles in s ynaptic differentiation of 'phasic'
and 'tonic' terminals. Previous studies have indicated that thzre are differences in
mitochondrial content and metabolic activity in these two types of motor neurons
(Lnenicka et al., 1986; Nguyen and Atwood 1992a: Nguyen and Atwood. 1994).
Furthemore, glutamate levels are higher in 'tonic' nerve endings (Shupliakov et al., 1995).
This could possibly explain the capacity of these neurons for facilitation upon high-
frequency stimulation by mechanisms which may involve making progressively more
glutamate-containing synaptic vesicles available for release. The higner quantal content of
the 'phasic' motor neuron cannot be explained by a higher glutamate content, since the
glutamate concentration is acniaily lower in the 'phasic' tenninals.
-
Previous snidies by Cooper et al. (1995) have shown that a higher concentration of
free ca2+ ions is generated in hi&-output terminais than in low-output tenninals for the
crayfish opener motor neuron ('tonic' type) at low frequencies of stimulation. Although
only relative changes in ~ a ' + using a fluorescent indicator were measured, the authors
reported that the high-output tenninals had relatively larger changes in fluorescence during
stimultation. Whether the differences in ~ a ' + signals are due to differences in the number
of ca2+ channeIs in the presynaptic tenninals or differences in the intratermïnal ca2+
handling or extrusion remains to be seen.
In surnmary, molecular differences other than the glutamate content are likely to
play an important role in determining the release properties of the 'phasic' and 'tonic'
motor neuron. Neuronal calcium binding proteins could be potential candidate molecules.
Frequenin: A neuronal calcium-binding protein
The neuronal calcium-binding protein frequenin has been implicated in the
modulation of synaptic efficacy (Pongs et al., 1993; Olaffson et al., 1995). Frequenin was
onginally discovered in a mutant in Drosophila. This mutant fly (V7) overexpresses
frequenin as a consequence of a chromosomal remangement.
-
At the protein level, frequenin consisrs of four structural alpha helical elements, the
so-called EF-hands, which are highiy conserved among the different members in the
superfamily of calcium-binding proteins. These "EF-hands" are the structural hailmarks
for calcium binding and were fint described in the calcium-binding protein calmodulin
(Nakayama et ai. 1994). Drosophila frequenin has k e n shown to bind calcium and to shift
its mobility upon calcium-binding (Pongs et ai., 1993).
Another stnicural feanire of this class of neuronal calcium-binding proteins is their
N-terminal myristoylation via an N-terminal glycine-residue. Studies of recoverin, a
related member of the calcium-binding protein superfamily, have shown that
myristoylation of recoverin decreases the affinity of calcium-binding, but introduces CO-
operativity of calcium binding (for review see Ames et al., 1996). Whether a sirnilar
mechanism is involved with frequenin remains to be elucidated. At the structural level,
studies of recoverin have shown that myristoylation provides a conformational switch (a
"myristoyl switch" ) which could be of importance for membrane-targeting and protein-
protein interaction (Ames et al., 1996). Knowing the rate constants and afinity of
frequenin for calcium wiIl undoubtedly be helpful in correlating its binding characteristics
to the calcium concentration measured in presynaptic terminais. Such studies will help to
provide the basis for understanding the structure-function relationships of frequenin.
-
Studies in Drosophila have shown that the overexpression of frequenin leads to an
enhancement of trammitter release under conditions of paired-pulse stimulation (Pongs et
al., 1993). Studies in Xenopus nerve-muscle cocultures have provided evidence that the
infusion of Xenopus frequenin into motor neurons not only enhances frequency-dependent
neurotransmitter release, but also alters the basal release properties (Olaffson et al., 1995).
The authon showed that the infusion of the protein increased the frequency of spontaneous
(quantal) synaptic currents, as well as their size. These functional studies predict a
strategic localization of frequenin to the sites of truismitter release.
Frequenin expression has been localized to the developing nervous system of
mammals both by in-situ hybridization and immunocytochemical means (OlafTson et al.,
1997). Recent cellular and sub-cellular locdization studies in mamrnalian brains have
demonstrated that frequenin is expressed primady in neurons and presumptive astrocytes.
In cultured hippocarnpal neurons, frequenin has been CO-locaiized with the dendritic
marker MAP-2. and the synaptic vesicle marker SV2 (Olaffson et al.. 1997). AIthough
these authors have shown that frequenin is primarily localized to synaptic compartments,
they have also found that frequenin is present in post-synaptic densities. The Functional
significance of this observation requires further investigation. Although the modulation of
-
transmitter release by frequenin has k e n documented, the rnolecular mechanisms for this
modulation are still unclear.
Frequenin has initiaily been described as a regulatot of guanylate cyclase (Pongs et
al. 1993). Activation of frequenin upon calcium-binding could therefore potentially
stimulate cGMP-dependent signaling pathways. Recent evidence suggested that the
mammalian isoform of frequenin, called neuronal-calcium-sensor- 1 (NCS- 1 ). can
substitute for calmodulin in Paramecium in vivo and in vitro (Schaad et al., 1996).
Furthemore, these authors have observed that non-myristoylated recombinant mouse
frequenin directly activates two Ca2+/calrnodulin-dependent enzymes. 3'-S'cyclic
nucleotide phosphodiesterase, and protein phosphatases such as calcineurin in vitro. Its
exact downstrearn signaling pathways, however, rernain to be investigaied.
Rivosecchi et al. (1994) proposed that a frequenin-dependent regulation of the
presynaptic Na'- ca2+ exchanger could account for the enhancement of facilitation in the
frequenin-overexpressing flies. Although the exact molecular bais for this modulation
remains to be investigated, these authors suggest that a stimulation of the exchanger via
cGMP is responsible for this effect.
-
In summary, studies so far have consistently suggested a role for frequenin in the
modulation of synaptic efficacy although the molecular mechanism for this modulation
remains to be examined.
Objective of the present study :
Preliminary studies have indicated a differential expression of frequenin in phasic vs. tonic
motor neuron terminals (Atwood et al. 1995). This was shown using a polyclonal
antiserum raised against DrosophiZu frequenin. The major objective of this thesis was to
establish the identity of molecules in the crustacean nervous systern which are responsible
for the immuno-reaction of the nerve terminais. In order to define precisely the molecules
present in crustacean motor neurons, 1 have further extended the original study and have
cloned the cDNA sequences for crayfkh and lobster frequenin. The crustacean frequenins
were identified as closely related members of the farnily of invertebrate frequenins.
In addition, the localization of frequenin in motor neuron terminals was further
investigated by direct CO-labeling studies. For this purpose, cross-reacting antibodies
against Drosophila vesicle proteins were used in conjunction with the anti- Drosophila
frequenin antibody. These CO-labeling studies revealed that frequenin is heavily expressed
in 'phasic', but less strongly in 'tonic' terminals.
-
Materiais & Methods :
Chernicals :
If not otherwise specified, laboratory chernicals of the highest grade available (Sigma)
were used for d l experiments.
Animais :
The spiny lobster nerve cord library was prepated from the nerve cord, including the
brain, of Panulirus interruptus, while the crayfish nerve cord library was prepared from
Procamam clarkii. For the immunocytochemistry, crayfish species frocambarus clarkii
and fmit flies species Drosophila rnelanogoster Canton S strain were used.
Cloning of crayfish frequenin :
In order to clone frequenin-related sequences from lobster and crayfkh, 1 have first
isolated an internal DNA fragment by the polymerase-chain reaction (PCR) using
degenerate primers and then used these internal fragments of frequenin cDNAs for
screening of the crayfish and lobster nerve cord cDNA libraries under high-stringency
conditions as outlined (Fig. 1).
The cDNA libraries used for the cloning of crayfish and lobster frequenin were provided by
Greg Lnenicka (SUNY, Albany) and Wulf-D. Krenz (Salk Institute, San Diego)
respectively.
-
a.) PCR ampl i fdon of an interna1 DNA fragment of frequenin :
5 ul of the cDNA Iibrary and degenerate primers against the conserved EF-han& of
frequenin were used for the PCR. As a given amino acid can be encoded by more than one
codon (degeneracy of the genetic code). primes were designed that Vary in the third
position of the codon (degenerate primer) to account for ail possible codons that code for
each arnino acid. The sequence of the primers used were Frq F1 5' PLAR ATH TAY AAR
CAR 'TTY 'ITY CC 3' and Frq B 1 5' CCR TCB 'ITR TI'Y TCR TCR AA 3'. The fonvard
primer Frq F1 anneais shortly 3' to the first EF-hand, while the backward primer Frq B 1
anneals 3'- to the third EF-hand of frequenin (Fig. 2) .
These primers were synthesized by the automatic oligo-synthesis service of the Samuel
Lunenfeld Research Institute of the Mount Sinai Hospital, Toronto.
The PCR was performed in a 50 ul volume consisting of :
5 ul nerve cord phage cDNA library
2 ul of each primer Frq FI and Frq B 1
1 ul of 10 rnM PCR nucleotide rnix (Boehnnger)
5 ul of 1Ox reaction buffer (Promega)
36.5 ul of sterile H 2 0 and
0.5 ul (corresponding to 2.5 units) of Taq Polymerase
(Promega).
-
The PCR reagents were mixed on ice, and then the Taq Polymerase was added. The
reaction was overlaid with minerai oil (Sigma), and a 'hot-start' in the Perkin-Elmer PCR
machine was performed.
For amplification of the frequenin cDNA fragment from the crayfkh and lobster
nerve cord libraries, the following profile was used :
initial denaturation at 94 C for 3 min,
94 C for 1 min, 45 C for 1 min and 72 C for 1.5 min for 35 cycles fcllowed by
a final extension step at 72 C for min.
25 ul of the PCR reaction was run on a 1.8 % agarose gel (Life Technologies). The
fragment of the expected size of 240 bp was excised from the gel. purified using the
GeneClean kit (BiolOl) and subcloned into the pCR II vector (Invitrogen). The insert-
containing DNA was purified using the Quiagen Mini-Prep kit (Quiagen) and subjected to
automatic di-deoxy-sequencing ( Biotechnology Service Centre, Banting Institute,
University of Toronto).
In the following, the cloning of crayfish frequenin will be described in more detail as
this is the major focus of the thesis. The cloning of the lobster frequenin followed identical
procedures : an interna1 fragment of the frequenin cDNA was fint amplified by PCR using
the degenerate prirners Frq FI and Frq B 1. This amplified fragment was then used to screen
the lobster nerve-cord cDNA library under high-stringenc y conditions and subsequent
procedures were identical to those described for crayfish frequenin.
-
Fig. 1 Schematic diagram of the procedure used for the isolation of the crayfish and lobster
frequenin full-length cDNA clones
-
Fig. 1
PCR amplification of the frequenin cDNA fragment
J+
High stringency screening of the nerve-cord cDNA library
1
Isolation and sequencing of a full-length frequenin cDNA clone
-
b.) Screening of the crayfish newe cord aerve cord cDNA iibrary :
1.) Plating of the library and trader to membranes
The Iibrary was plated on C600 bacteria (Invitrogen) at about lo5 plaques per large
(15 cm in diameter) Petri dish. The plates were incubated at 37 OC for 8 hrs until the
plaques had reached about I mm in diameter and the plates were subsequently chilled at 4
OC.
Each piate was transferred to Hybond N+ membranes (Arnersham) according to the
manufacturer's instructions.The membrane of corresponding size was placed on the agar
surface. Afier 1 min the membrane was removed and placed colony side up on a pad of
absorbent filter paper (Whatman) soaked in denaturing solution. After 7 min, the
membrane was transferred to a pad of absorbent filter paper (Whatman) soaked with
neutraiizing solution, and then after 3 min transferred to a fresh pad soaked with the same
solution. The filter was washed in 2 x SSC (sodium chloridel sodium citrate 20x : 3M
NaCl, 0.3 M Nqcitrate 2Hz0, the pH was adjusted to 7.0 with 1 M HCI), transfered to dry
filter paper and air-dried or dried at 80 OC in an oven for 10 min (Ausubel et al., 1987).
II.) Random-primed labeling of the interna1 PCR fragment
Labeled frequenin DNA for the high-stringency screening of the nerve cord libraries
was prepared using the random primed DNA labeiing kit (Boehringer). This protocol of
random primed DNA labeling developed by Feinberg and Vogelstein (1994) is based on
the hybridization of a mixture of al1 possible hexanucleotides to the DNA to be labeled.
-
To carry out the labeling reaction, the purified frequenin cDNA fragment was
denatured by boiling resulting in iinear DNA molecules. These were then annealed to
random-sequence hexanucleotides and incubated with the Klenow fragment (the large
fragment of DNA polyrnerase 1, lacking 5'+3' exonuclease activity) in the presence of
nucleotides, including a 32~-labeled nucleotide. in this way, the hexanucleotides prime the
frequenin DNA at various positions dong the template and are extended to generate
double-stranded DNA that is unifomly labeled on both strands.
-
Fig. 2 Positions of the two degenerate primers Frq FI and B 1 used for PCR amplification
of the crayfish and lobster cDNA fragments. The protein sequence of Drosophila frequenin
with its EF-hands in bold are shown. The positions where the two primers anneal are
underlined.
-
Figure 2. Protein sequence of Drosophila frequenin (Pongs et al.. 1993). The EF- hands are shown in bold and the regions where the primers anneal are underlined.
-
According to the protocol, 25-50 ng of frequenin DNA was denatured by heating for 10 min
at 100 OC and subsequent cooling on ice. The following reagents were then added to a
microtube on ice :
25-50 ng denatured DNA
3 u1 dATP,dGTP and dTTP
2 ul reaction mixture
5 ul = 50 uCi d m , 3000 Ci/mmol, aequeous solution.
ddH$ to make up the reaction volume to 19 ul and
1 ul Klenow Enzyme (4 u/ul).
The reaction was then incubated for 30 min at 37 OC and stopped by adding 2 ul 0.2 M
EDTA (pH 8.0) This reaction mixture was then passed over a Sephadex 50 colurnn
(Boehringer) to separate the labeled DNA from the unincorporated nucleotides. The
arnount of radioactivity was then determined by scintillation counting : 3-5 106 cpm of the
labeled probe were added to each hybridization tube in order to standardize the arnount of
radioactive probe added.
III.) High-stringency hybridization in aqueous solution
In order to identiQ DNA sequences that match the sequences of the frequenin cDNA
fragment, I screened the crayfish nerve-cord library transferred to membranes as described
-
in 1.) under high stringency conditions in aqueous solution as described by Church &
Gilbert (1984). Ail steps were carried out at 65 OC in a rotating hybridization oven.
The membranes were prehybridized in hybridization solution II (1 % crystalline
solution BSA, 1 m M EDTA (Fluka), 7 % SDS (Boehnnger), 0.5 M NaHP04, pH 7.2 ,)
for at Ieast 1 hr. 3-5 mill. cpm of the labeled probe was then added and the hybridization
was carried out for at least 16 hr. The hybrïdzation solution was then poured out and the
membranes were washed twice with low-stringency wash buffer II ( 0.5 % BSA, 1 rnM
Na2EDTA, 5 % SDS (Boehnnger), 40 mM NaHP04, pH 7.2) followed by multiple (5-8)
quick washes and a final wash for 20 min in high-stringency wash buffer II ( 1 8 SDS
(Boehnnger), L mM Na?EDTA, 40 rnM NaHP04, pH 7.2). The membranes were then
removed from the hybridization tubes, briefly ciried on filter paper (Whatman) and exposed
to an autoradiognphic film (Kodak X-Omat) for 16-24 hn. Areas of the plates
corresponding to areas of hybridization signs were eluted from the corresponding plate and
replated at lower density until a single plaque could be identified. Single plaques were then
subcloned into the pCR vector (Invitrogen) and the largest insert (1.5 kb) subjected to
automatic di-deoxysequencing (Biotechnology Service Centre, Banting Institute,
University of Toronto).
-
Immunocytochemistry :
a.) Animais :
1.) Crayfiih :
Freshwater crayfish, Procambarus clarkii Girard, of intermediate size ( 6 cm ) from
rostrum to telson, mass 5-6g) were obtained from Atchafdaya Biologicd Supply Co..
Louisana, USA, and maintained comrnunally in dechlorinated aerated tapwater at 16 OC on
a 12 hr light/dark cycle.
The animaIsT diet consisted of lentils, carrots and fishfood (Tetramin).
II.) Drosophila mehnogasfer :
Drosophih rnelanogaster larvae (Canton-S wild-type) were reared on standard
cornmeal
medium (Stewart et ai, 1995). Wandering 3rd instar larvae were used for al1 experiments.
b.) Preparations :
1.) Crayfsh muscle dissections :
For the dissection of the abdominal extensor muscles, the shell was cut through dong
the lower lateral border of the abdomen on each side. The lower half of the abdomen
together with the flexor musculature was then readily separated from the upper half of the
abdomen and the extensor musculature. The preparation was then anchored with the
ventral (muscle) side up by puaing a pin through the shell at each end into a Sylgard-dish.
The preparation was kept in crayfish solution and processed as described below. This
-
preparation of the abdominal extensor muscles is described in more detail in (Atwood and
Pamas, 1966).
For the dissection of the leg extensor muscle, the first (or occasionally the second) pair
of walking legs was used. The muscle was exposed by removing the cuticle on the lateral
aspect of the meropodite dong with the entire flexor muscle and also the main leg nerve.
The preparation was then superfked with crayfi~sh solution until pre-fixation.
For preparation of the abdominal slow flexor muscle. the abdomen was pinned ventral-
side up in a Sylgard dissecting dish and bathed in crayfkh solution. The superficial
exoskeleton was then carefully removed from the abdominal segments (1-5) exposing the
slow (superficial) flexor muscles (Kennedy and Takeda, 1965 b).
II.) Preparation of Drosophihz 3rd instar larval fillets :
3rd instar iarvae were dissected by making a longitudinal mid-dorsal incision and
pinning the cuticle flat using insect pins as described in Stewart et al. (1996). The internai
organs were carefully removed to expose body-wall muscles and the nervous system. These
3rd instar fillets were then fixed in the dish and processed for staining as described below
for the crayfish muscle specimens.
c.) Solutions :
1.) Crayfsh saline :
The composition of the crayfish physiological solution used is as follows (concentrations
in rnM; modified after van Harreveld, 1936) : NaCI 205.3, KCl 5.3, CaClz 2H20 13.5,
MgCl 6Hr0 2.5 and Hepes buffer, 0.5 mM, adjusted to pH 7.4.
-
The different crayfish muscle preparations were pre-fixed in 3 % paraformaldehyde
(PFA) and 4 % sucrose in phosphate-buffered saline (PBS) containing (in mM) : 137
NaCl, 2.7 K I , 4.3 Na2HPQ 7H20, 1.4 KH2P04, pH - 7.3.
Il.) Solution used for Drosophih Iarval dissection :
The third instar larval fillet dissections were performed in Schneiders medium (Life
Technologies).
d.) Antibodies:
The anti-Drosophiiu frequenin antibody was directed against the full-length Drosophiia
frequenin (Pongs et ai., 1993). It recognizes a 2 1 kDa protein in crayfish ganglia and also
cross-reacts with the recently-cloned crayfish frequenin (see Appendix). The antibody was
a gift from O. Pongs (Hamburg, Gerrnany) and was used at a dilution of 1: 100. The anti-
Drosophila synapsin antibody (SYNORFI) was used at a dilution of 1:20 and was the
generous gifi of E. Buchner (Klagges et al., 1995). The anti-Drosophila dynamin antibody
(Estes et al. 1996) was a gift of R. Kelly and J. Roos and was used at a dilution of 1: 100.
The anti-Drosophila cysteine-string protein antibody was kindly provided by K. Zinsmaier
and used at 1 5 (Zinsmaier et ai., 1994). It showed only weak immunoreactivity and was
not characterized further. The anti-Drosophila synaptotagmin was provided by H. Bellen
(see Cooper et al., 1995 a). This antibody was raised in rabbits and therefore was not used
for the double-labeling experiments with the anti-Drosophila frequenin antibody which
was also raised in rabbits. A second anti-synapsin antibody ((3-304) was donated by A.
-
Czernik and P. Greengard and used at 1 : 100. It produced a staining similar to the one using
the anti-Drosophila synapsin (SYNORFI) mouse monoclonal antibody (Klagges et al.,
1995). For the double-labeling shown here however, oniy the mouse monclonai anti-
synapsin antibody was used.
Except for the mti-Drosophila synapsin monoclonal antibody which was used as
hybridoma supernatant. al1 polyclonal antisera were used as aKinity-purified antibodies.
e.) Fixation and staining :
1. Prefuration of crayfih muscle preparations :
The craflsh nerve muscle preparations were prefixed in 3 9% PFA and 4 % sucrose in
PBS.
After this prefixation, which lasted 30 min at room temperature, the specimens were
transferred briefly to the crayfish standard solution. Then, thin layers of the laterai parts of
the (deep) abdominal extensor muscles were sliced off the surface with a razor blade
scalpel. After prefixation for 30 min at room temperature, the leg exensor muscle and the
slow flexor muscle were also transferred bnefly to the standard crayfish solution, separated
from the exoskeleton, and then cut into smaller thin pieces. These 3 different nerve-muscle
preparations were then further fixed and processed as described below.
ïI. Fixation and sbining :
Crayfish nerve-muscle preparations, prefixed and dissected into smail pieces as
described above, were fixed in PBS solution containing 3% paraformaldehyde and 4%
sucrose for an additionai 1 hr at room temperature.
-
The 3rd instar larvd fillets were only fixed for 1 hr in PBS containing 3%
paraformaldehyde and 4% sucrose. No further dissection was necessary for this
preparation..
The fixed specimens (both the Drosophila larvd fillets and the crayfish nerve-muscle
preparations) were then transfered to individual wells of a multi-well plate (Nunc), washed
three times at one-hour intervals with PBS, and permeabilized in PBS containing 2 %
normal goat serum (NGS) (Jackson Imrnunochernicds), 2% bovine serum albumin (BSA)
(Sigma), and 4 % Triton-X-100 (Sigma) (hence called 4% PBT) for one hour at 4 OC.
These preparations were incubated with the prïmary antibodies in 0.4% PBT ovemight.
Following incubation in the primary antibody, the specimens were washed three times as
above and incubated with secondary antibodies in 0.4% PBT for one hour. Secondary
antibodies were then removed and specimens washed three times in PBS and mounted in
Permafluor (Immunon) mounting medium.
Anti-rabbit or anti-mouse secondary antibodies were obtained from Jackson
Immunolabs and were used at 1:40 dilution in 0.4 % PBT. For the double-labeling
expenments in which both frequenin and synapsin immunoreactivity were demonstrated ,
goat anti-mouse-cy3- and donkey anti-rabbitcy5sonjugated secondary antibodies (Jackson
Immunochemicals) were used. The goat anti-mouse cy3-conjugated secondary antibody
was used for the mouse monoclonal anti-Drosophila spapsin antibody (SYNORF I ), while
the donkey anti-rabbit cy5-conjugated secondary antibody was used for the polyclonal anti-
Drosophila frequenin antibody raised in rabbits. Cy3 and cy5 are both cyanine dyes. Cy5
has been used with a variety of other fluorophores due to the wide separation of its
-
emission spectrum from that of shorther-wavelength-emitting fluorophores. Cy5 is excited
maximally nzar 650 nrn and fluorexences maximally near 670 nm. Cy3 can be maximally
excited near 550 nm with a peak fluorescence near 570 nm.
m., Imagïng :
The mounted specimens were imagined with a MRC 600 confocal microscope (Bio-
Rad, Missaussaga, ON) equipped with a Krypton/ Argon laser using either a 40x or 63x
oil-immersion objective and the COMOS software. For excitation of cy3-conjugated
secondary antibodies, the YHS filter block was used, while for the excitation of cy5-
conjugated secondary antibodies the RBS filter block was used. The YHS filter block
consists of the 568 nm DFIO excitor filter, the dichroic reflector 585 DRLP and the
emission filter 585 EFLP. The REiS filter block consists of the excitor filter 647 DFIO, the
dichroic reflector 660 DRLP and the emission filter 680 EF32.
To ensure confocality, the pinhole was set to - 0.5 in al1 experiments. If z-scans were taken, the distance between individual planes was set to 0.5 um. The individual z-scans
were projected and saved into a single projected image file. These projected z-scans were
then imported into Confocal Assistant (Bio-Rad) and Adobe Photoshop for generation of
the red-green overlays shown in the figures.
-
1 have cloned frequenin cDNAs from crayfi~sh and lobster nerve cord libraries and
andyzed the primary sequences coded by these cDNAs. Antibodies against two synaptic
proteins, synapsin and dynamin, were used to reveal the localization of synapses in
different crayfish nerve-muscle preparations and to compare these with the location of the
sites of expression of frequenin imrnunoreactivity, as revealed by the anti-Drosophila
frequenin antibody.
a.) Cloning and analysis of crayfiih and lobster frequenin cDNAs :
Crayfish frequenin was cloned from a nerve-cord library using a combination of PCR
and high-suingency screening (Fig. 1). As the EF-hands at the conserved regions were
considered to be conserved among members of the frequenin family, two degenerate
pnmers (Frq FI and BI ) was designed to amplify a fragment from the nerve cord library
(Fig. 2). This crayfish frequenin cDNA fragment was then used to identify a full-length
frequenin cDNA clone using the procedures described in the "Material and Methods"
section (Fig. 1).
-
FoIlowing identical procedures to those described for crayfish frequenin, I amplified an
intemal lobster fkquenin cDNA fragment from the lobster nerve cord library (Fig. 4) and
used it to screen the lobster nerve cord cDNA library under high stringency conditions.
A PCR fragment of the expected size was arnplified from the lobster nerve cord (Fig.
4) and crayfish nerve cord cDNA library (Fig. 3) using the primen Frq FI and B 1. These
fragments were used for high-stringency screening of the corresponding library in order to
isolate a full-length lobster and crayfish frequenin cDNA clone.
The cra*sh cDNA frequenin clone with the largest insert of 1.5 kb identified (Fig. 5)
was subjected to automatic di-deoxysequencing. The sequence obtained is shown in Fig. 6 .
The analysis of al1 possible open reading frames (a,b and c) of the crayfish cDNA
sequence was performed using the MAP program (GCG package. Wisconsin, Madison).
MAP translates the DNA sequences into al1 possible open reading frames ((a), (b) and (c)).
The Iongest open reading frame for craytiish frequenin that could be identified by MAP is
reading frarne (a). This reading frarne starts at position 204 (marked by an arrow) and ends
at position 797 (Fig. 6). The crayfish frequenin open reading frarne is preceded and
followed by several stop codons demonstrating that this open-reading frame is a full-length
open reading frame (Fig. 6) .
-
Fig. 3 Gel electrophoresis on a 1.8 % agarose gel of the PCR products amplified from the
crayfish nervetord library using the degenerate primers Frq F1 and Frq B 1. The PCR
fragment of the expected size (at the arrow) was excised and subcloned.
A: 1 kb ladder rnolecular weight marker (Life Technologies), B : First amplification from
the crayfi~sh nerve-cord library, C : 10 ul of the first amplification reaction was amplified
for a second time, D : PCR amplification from Drosophila frequenin (positive control)
using the Frq FI and Frq B 1 primers.
-
Fig. 4 Gel electrophoresis on a 1.8 % agarose gel of the PCR products amplified from the
lobster nervetord library using the degenerate primers Frq FI and Frq BI. The PCR
fragment of the expected size (at the arrow) was excised and subcloned.
A: 1 kb ladder molecular weight marker (Life Technologies), B : First amplification from
the lobster nerve-cord library, C : 10 ul of the fint amplification reaction was amplified for
a second tirne, D : PCR amplification from Drosophila frequenin (positive control) using
the Frq F 1 and Frq B 1 primers.
-
Fig 4
-
Reading frame (a) encodes for a protein of 193 amino acids (Fig. 8).The extracted
protein sequence of crayfish frequenin w ith its highlighted structurai features, the EF- hands
and the possible N-rnyristoylation motif, is shown in Fig. 8.
The andysis of the lobster frequenin cDNA clone identified with the largest insert
(1.2 kb) (Fig. 5) revealed, using MAP (GCG package, Wisconsin, Madison), a single open
reading frame starting at position 58 and terminating at position 610 of reading frame (a)
(Fig. 7). This open reading frarne is also preceeded and followed by stop codons (Fig. 7).
This longest open-reading frarne that could be identified encodes for protein of 188 arnino
acids in length (Fig. 9). The extracted protein sequence with its structural features, the EF-
hands and the N-terminal glycine residue that could serve as a myristoylation site, is shown
in Fig. 9.
A pairwise cornpanion of crayfish, lobster and Drosophila frequenin cDNA sequences
using GAP (GCG software package) revealed that the three sequences are about 67 %
identical to each other. GAP aligns two sequences in their entirety by introducing gaps into
the sequence. Gaps are introduced into each of the two sequences to be digned, until a
maximum of sequence sirnilarity is reached. At the protein level, a painvise alignent using
GAP showed that the arnino acid sequences of crayfish and Iobster frequenin are about 74
-
and 80 % identical to that of Drosophika frequenin, respectively (Fig. 10 and I l ) . Lobster
and crayf~sh frequenin share 74 % identity at the arnino acid Ievel (Fig. 12).
-
Fig. 5 Analysis of the insert-size of purified single plaques of crayfi~sh and Iobster
frequenin cDNAs identified from the high-stringency screen by electrophoresis on a 1.2 %
gel.
A: 1 kb ladder molecular weight marker (Life Technolgies) , B: insert from a single plaque
identified from the crayfi~sh nerve-cord library. C : insert from an independent plaque
identified from the crayfikh nerve-cord library. D : insen frorn a single plaque identified
from the lobster neme-cord library. E : insert from a single plaque identified from the
crayfish nerve-cord library. This insert is identical to B. as it was denved from the same
plaque.
The longest inserts frorn each screen, (B) from the crayfkh nerve-cord library and (D) from
the lobster nerve-cord library, were subjected to sequencing.
-
Fig. 6 Translation of the crayfish frequenin cDNA into the possible open reading frmes
using MAP (GCG software package) This program translates the nucleic acid sequence in
any of the possible translation frames. The three possible reading reading frarnes are shown
as (a), (b) and (c). The translation intiation site, the initial methionine, of the crayfish
frequenin is marked by an m o w (position 204). The open reading frarne (a) of the crayfkh
frequenin ends at position 797 as depicted by an asterisk.
-
A E R S G S R Q E T Q R Y Y I N K R G E - Q R E A V A D K R R R D I I * T K G V K -
R E K R * Q T R D A E I L Y K Q K G * S -
a A S V I E Y H D L D E S C P N H L S S G - b H P L L N T M I W M N H A P T T S V V V - c I R Y * I P * S G * I M P Q P P Q * W C -
a V S L C f F L P S P P L S S P L P S S P - b " V C A N S S L L L L F H H L S L P R P - c * E S V L I P P F S S S F I T S P F L A Q -
- R R H Q P P A N M G K K N S K L K Q E T -
D V T N H L Q T W E R R T P N S N R K L - T S P T T C K H G K E E L Q T Q T G N Y -
I Q K L C E E T Y F S D K E I K Q W H K - S R N C V K K L I F L T K K S S N G T S -
P E T V * R N L F F * Q R N Q A M A Q A -
~ K I Y K Q F F P Q G D P T K V A S L V - S R Y T S S S S P R E T P P R L P L W C -
-
F R V F D E N N D G A I E F E E F I R A S E Y L M R T M M V R S S L R S S S G R
Q S I * * E Q * W C D R V * G V H Q G A
L S I T S R G N V D E K L L W A F R L Y S L S H P E E M S T R N S C G R S G Y T
L Y H I Q R K C R R E T P V G V P A I R
D V D N D G F I T R E E M Y S I V D A I T S I T M A L S P E R K C T A * L T L Y
R R * R W L Y H Q R G N V Q H S * R Y I
TACCAGATGGT~CAGGCTCCTGAGGCAGAGGATGAGAACACGCCACAGAAACGCGTC ---------+-.--.----+---------+--------+---------+---------+ A T G G T C T A C C A C C C C G T C C G A G G A C T C C G T C T C C T A C T C C
Y Q M V G Q A P E A E D E N T P Q K R V T R W W G R L L R Q R M R T R H R N A S
P O G G A G S * G R G * E H A T E T R R
D K I F S Q M R R G S N S F A F Y P G R 1 . R S S R R C E G V R I L L L F T L E E * D L L A D A K G F E F F C F L P W K K
N T H K S R I P A H W R P L L V D P S S I L I K A E F Q H T G G R Y * W I R A R Y S * K P N S S T L A A V T S G S E L G
V P S t A * S W S * L F P V * K L L S A Y Q A W R N H G H S C F L C E N C Y P L
T K L G V I M V I A V S C V K I V I R S
CACAATTCCACACAACATACGAGCCGGAAGCATAAAGTGTWGCC - - - - - - - - - + - - - - - - - - - + - - - - - - - - - + - - - - - - - - - + - - - - - - 886
-
Fig. 7 Translation of the lobster frequenin cDNA into the possible open reading frarnes
using MAP (GCG software package) This program translates the nucleic acid sequence in
any of the possible translation frames. The three possible op& reading frames are shown as
(a), (b) and (c). The translation intiation site, the initial methionine, of the lobster frequenin
is marked by an arrow (position 58). The open reading frame (a) of the lobster frequenin
ends at position 688 as depicted by an asterisk.
-
Fig. 7 1
K T N Y R G I Q E K Q T S L N I L V T D - R Q T T E E F K R S K R P * I F k L P T -
D K L Q R N S R E A N V P E Y F S Y R Q -
E T G F I K I Y K Q F F P Q G D P T K F K R A S S R S I N S S S H K G T P P S S
N G L H Q D L f T V L P T R G P H Q V R
A S L V F R V F D E N N D G S I E F E E R R L S S E S L T K I M T G Q S S S R N
V A C L P S L * R K * * R V N R V R G I
F I R A L S I T S R G N V D E K L . L ' W A S S E R S P S H L E G T W M K S Y Y - G L
H P S A L H H I k R E R G * K A T M G F
F K L Y D V D N D G F I T R E E M Y S I S S C M T * T T M V S S P E R R C I V f
Q A V * R R Q R W F H H P R G D V * Y S
-
V D A I Y Q M V G H A P E A A D E N T P - W M P V T R W S V M P Q K L Q M R T H H -
G C H I P D G R S C P R S C R * E H T T -
Q K R V D K I F S Q M D K N H D E K L T - R R G S I R S S H K W T R I M T R N L L -
E E G R * D L L T N G Q E S * R E T Y F -
L E E F K E G S N A D P R I V Q A L S L - W R S S R R G P M P T Q E L C R R C P L -
G G V Q G G V Q C R P K N C A G A V P W -
G D N * E P T H H D V H K V R T T E W E - V T T K N L L I M M F I K F V L P S G K -
* Q L R T Y S S * C S * S S Y Y R V G K -
N G T S L A L R V * G * I M I C * H C S - M E P P * H Y E Y R V R L * F A N T A A -
W N L L S I T S I G L D Y D L L T L Q P -
L L S S C L T - S C P L V * -
L V L L F D -
-
As in Drosophila frequenin (Pongs et al.. 1993), the first and fourth EF-hand are most
likely non-functional based on the analysis of the prirnary structure. In the first EF-hand of
crayfkh and lobster frequenin, the CPNG amino acid motif is found (Fig. 8 and 9,
respectively). The cysteine and proline residues are believed to impair the a-helical
arrangement of this EF-hand (Ames et ai., 1996). The fourth EF-hand of crayfïsh and
lobster contains a number of charged amino acid residues as does the fourth EF-hand of
Drosophila frequenin. The third EF-han& are aimost 100 % identical between crayfish,
lobster and Drosophila frequenins except for a single amino acid change (Y+ F) in third
EF-hand of both crayfish and lobster frequenins (Figs. 10 and 1 1). The second EF-hand is
100 % conserved in lobster frequenin compared to Drosophila frequenin (Fig. 1 1) A
single arnino acid substitution has occurred in the second EF-hands of crayfish frequenin
(A+S) when compared to Drosophila frequenin (Fig. 10).
The N-terminal myristoylation consensus sequence (MGXXS) is also found in crayfish
frequenin (Fig. 8). whereas this motif is absent in lobster frequenin (Fig. 9). Whether the
glycine-residue imrnediately downstream (YRGIQ) (Fig. 9) can function as myristoylation
site in lobster frequenin remains to be determined.
It is apparent from the sequence cornparison that the lowest degree of sequence
conservation is found in the N- and C-terminus (Fig. 11 and 12). While the N-terminus of
-
craytTsh and Drosophila frequenin is rather similar, the C-terminus is not (Fig. 1 1 . ) . The
opposite pattern of sequence conservation is found for lobster and Drosophila frequenin
(Fig. I l ) . Between these two proteins, the C-terminus is alrnost identical except for 5
amino acid changes, while the N-terminus is less well conserved.
-
Fig 8 Amino acid sequence of craflsh frequenin, as extracted by MAP (GCG software
package). The EF-hands are shown in bold, the N-terminal glycine residue that might be
myristoylated is also shown in bold and marked by an asterisk.
-
Fig. 8
Crayfish Frequenin
PSSVPSLA* 193
-
Fig. 9 Amino acid sequence of lobster frequenin, as extracted by MAP (GCG software
package).
The EF-hands are shown in bold, the N-terminal glycine residue that rnight be
myristoylated is also shown in bold and marked by an asterisk.
-
Fig. 9
Lobster Frequenin
-
Fig. 10 Pair-wise alignment of crayfish and Drosophila frequenin protein sequences using
GAP (GCG software package). The protein sequence of c-sh frequenin is shown in the
upper and Drosophila frequenin is lower panel. GAP generates the best dignment of two
sequences in their entirety by introducing gaps in either of the sequences to effect better
alignment. The parameters used for each aiignment and the percent identity and similarity
are shown above each sequence alignment. Identical arnino acids are marked by lines,
while sirnilar arnino acids are marked by dots. The four EF-hands are marked by the
arrows.
-
Fig. 10
Pairwise-alignment of crayt'îsh and Drosophila frequenin protein sequences
Craflsh (upper panel) vs. Drosophüa frequenin (lower panel)
Ga Weight: 3.000 1 Average Match: 0.540 Lengt Weight : 0.100 Average Mismatch: -0 -396 Quaiitv: 217.8 Length : 192 - ~ a t i 6 : 1.171 Gaps :
Percent Simi 1 a r i ty : 85.405 Percent Identi ty : 73.514
-
Fig. 1 1 Pair-wise dignment of lobster and Drosophila frequenin protein sequences using
GAP (GCG software package). The protein sequence of lobster frequenin is shown in the
upper and Drosophila frequenin is lower panel. GAP generates the best alignment of two
sequences in their entirety by introducing gaps in either of the sequences to effect better
alignment. The parameters used for each dignment and the percent identity and similarity
are shown above each sequence alignment. Identical amino acids are marked by lines,
while similar amino acids are marked by dots. The four EF-hands are makred by the
arrows.
-
Fig. 11
Pair-wise aügnment of lobster and Drosophila frequenin protein sequences
b.) Lobster (npper panel) vs. Drosophila frequenin (lower panel)
Ga Weight: 3.000 I: Average Match: 0 -540 Lengt Weight : 0; 100 Average Mismatch : -0.396 Quality: 229.9 Length: 190
Ratio: 1.249 Ga ps : 1. Percent Slmi 1 art ty : 89.011 Percent Identi ty : 80.769
-
Fig. 12 Pair-wise alignment of lobster and crayfkh frequenin protein sequences using
GAP ( K G software package). The protein sequence of lobster frequenin is shown in the
upper and crayfish frequenin is lower panel. GAP generates the best alignment of two
sequences in their entirety by introducing gaps in either of the sequences to effect better
dignment. The parameters used for each dignment and the percent identity and similarity
are shown above each sequence dignment. Identical amino acids are marked by lines,
while similar amino acids are marked by dots. The four EF-hands are marked by the
mows.
-
Fig. 12
Pair-wise alignment of crayfish and lobster frequenin protein sequences
Lobster (upper panel) vs. crayfish frequenin (lower panel)
Ga Weight: 3.000 i Average Match : 0.540 Lengt Weight: 0.100 Average Mismatch : -0 -396 Quality: 204.2 Length : 198
Ratio: 1.110 Gaps : 1 Percent Simi 1 a r i ty: 78.333 Percent Identi ty: 73 -889
. . ---- -- - 1 . . . . M ~ ~ I Q E K Q ~ L ~ I L ~ E I K ~ K G F L K D C P F I 46
1 O - 1 '#'&!#b~'&bk!&!k!i~b 50 . i MGKKMSKLKQ~~IQKLCEEPIFSDK K K L
-
b.) Immunolocaiization of synapsin, dynamin and frequenin-üke molecules at
crayfiih and Drosophilrr neuromuscdar junctions :
As descnbed in " Materials and Methods ", the antibodies used to investigate the
expression of these three proteins were originally generated against the Drosophila
isoforrns, but appear to cross-react with their crayfish counterparts (Table 1). This is in
agreement with a previous report using an anti-Drosophila synaptotagrnin antibody
(Cooper et al., 1995 a). These authors showed that synaptotagmin-like immunoreactivi ty
was present in both 'phasic' and 'tonic' terminais. For cornparison and to confirm the
specificity of the antibodies used, the localization of synapsin, dynamin and frequenin
immunoreactivity at the mature neuromuscular junction of Drosophila larvae was
exarnined (Fig. 16).
h the following, the antibodies tried and the observed pattern of immunoreactivity
are described. As the anti-Drosophila frequenin antibody is a polyclonai antibody raised in
rabbits (see "Materials and Methods"), on1 y the anti-Drosophila s ynapsin mouse
monoclonal antibody (SYNORFI) was further characterized and used for the double-
labeling studies. The mouse monoclonal anti-Drosophila synapsin antibody was used to
-
reveal synaptic tenninals, while the anti-Drosophiln frequenin antibody was used to reveal
the localization of fiequenin irnmunoreactivity in these direct CO-labeling experiments. The
same strategy was applied to compare the localization of synapsin and dynamin
irnrnunoreactivity. The anti-synapsin E2 domain antibody (Table 1) is polyclonal and
therefore was not further characterized. The anti-Drosophila cysteine-string protein (csp)
antibody shows only weak irnmunoreactivity in crayfkh and therefore was not used further
(Zinsmaier et al., 1994). Instead, the anti-dynamin antibody was used as another antibody
for labeling synaptic tenninals. The anti-Drosophila synaptotagmin antibody is also
polyclonal and therefore was not suitable for double-labeling experiments with the
polyclonal anti-Drosophila frequenin antibody.
In the following, one representative example for each of the three different crayfish
muscle preparation is shown. The staining shown represent an example of 6-8 different
experiments.
ce) Immunolocalization at crayfish neuromuscular junctions :
1.) Synapsin :
In the three different crayfish nerve-muscle preparations exmined, there was a good
immunoreactivity in the nerve terminals. This implies that a synapsin-like molecule was
expressed in the terminals of both 'phasic' and 'tonic' motor axons (Fig. 13 a. 14 a and 15 a
-
and 13 d, 14d and 15d). We have obtained an identical staining pattern using a polyclonal
antibody against the conserved E2 domain of synapsins (G-304, data not shown). The
expression pattern of synapsin irnmunoreactivity almost perfectly overlaps with that of
dynamin in al1 three nerve-muscle preparations exarnined (Fig. 13 c, 14 c and 15 c for the
overlay).
-
Table 1 Antibodies tested for cross-reactivity in crayfih
S ynapsin
E2 domain of synapsin
(G-304)
Dynarnin
S ynaptotagmin
Cysteine-string protein
Species Mono-/ Po fyclonul Cross-reacitiviîy in
c r ~ f i s h
Drosophila Mono- Yes
Mammals Poly- Yes
D rosop hila Poly-
Drczsophila Poly-
Drosoph ila Mono-
Yes
Yes
(Very weak)
-
II.) Dynamin :
In the abdominal slow flexor muscles, the fast abdominal extensor and the leg
extensor nerve-muscle prepration, the staining pattern of the anti-dynamin antibody
included both thin filiform and more varicose terminds (Fig. 13 b, 14 b and 15 b,
respectively). This implies that a dynamin-related molecule is present in crayfi~sh motor
nerve terminals. The expression pattern of the dynamin imrnunoreactivity is dmost
identical to that of synapsin immunoreactivity at these crustacean neuromuscular junctions
(Fig. 13 c, 14 c and 15 c for the overlay).
III) Frequenin :
In the crayfish fast abdominal extensor muscle, frequenin immunoreactivity was
detected in al1 thin filiform terminals (Fig. 15 e). Its expression pattern was almost identical
with that of synapsin (Fig. 14 d and f for the overlay).
In the abdominal slow flexor muscle, frequenin immunoreactivity was restricted to
a subpopulation of terminals (Fig 15 e and f for the overlay ). These terminals appeared to
be more varicose than the thin filiform terminals stained in the abdominal extensor muscle.
The same feanires were revealed by the anti-synapsin staining ( Fig. 15 d).
-
The dichotomy of frequenin irnmunoreactivity became more apparent in the Ieg
extensor muscle. In this nerve-muscle preparation, frequenin immunoreactivity was found
only in the thin filiform, but not the larger varicose terminais (Fig. 13 e and f for the
overlay). In particular, in regions where thin filiform and more varicose terminals run
close together as revealed by the anti- synapsin staining (Fig. 13 d), only the thin filiform
terminais strongly express frequenin immunoreactivity (Fig. 13 f for the overlay). Notably,
frequenin irnrnunoreactivity was found not to be expressed in al1 thin filiform terminals
(Fig. 13 e.). This implies that not al1 terminals of the same axon contain the same amount
of the immunoreactive molecule,
d.) Immunolocaüzation of synapsin, dynamin and frequenin at the Drosophila larval
neurornuscular junction :
1.) Synapsin :
At the Drosophila larval neuromuscular junction, spapsin is expressed in both types
of synaptic terminals (type lh for Mg and 1s for gndl) (Atwood et al., 1993) (Fig. 16 a
and d). The localization of synapsin is virtually identically to that of dynamin (Fig. 16 f for
the overlay).
-
II.) Dynamin :
Dynamin expression was detected in Ib and 1s terminals. Its expression pattern
overlapped with that of synapsin (Fig. 16 e and f for the overlay).
III.) Frequenin :
At the Iarvai neuromuscular junction, frequenin is expressed in both Ib and Ts
terminais (Fig. 16 b). Its expression pattern dmost completely ovelaps with the synapsin
expression pattern (Fig. 16 b and c for the overlay).
-
Fig. 13 Neuromuscular junction of the crayfish leg extensor muscle. A : synapsin
immunoreactivity (red); B : dynamin immnuoreactivity (green); C : the corresponding
overlay for A and B is shown in yellow. D : synapsin (red); E: frequenin (green) ; the
overlay of D and E is shown in F. Note that while synapsin (A,D) and dynamin (B) reveal
both thin filiform and more varicose endings, frequenin loclization is restricted to thin
filiform nerve endings (E and the overlay in F in yellow). A heterogeneity is revealed
among the thin filiform terminals, as some filiform endings shown in D are devoid of
frequenin immunoreactivity (see E and F). The size of the scale bar is 10 ym.
-
68
Fig 13
-
Fig. 14 Neuromuscular preparation of the crayfïsh fast extensor muscles of the abdomen.
Synapsin and dynamin immunoreactivity are reveaied in A (red) and B (green). Synapsin
and dynamin immunoreactivity reveals thin filiform nerve endings in a continous network
over the muscle (the overlay of A and B is shown in C in yellow). Frequenin
immunoreactivity is shown in E (green) which almost completel y CO-localizes with
corresponding regions as revealed by spapsin shown in D. The CO-localization of synapsin
(D) and frequenin (E) is shown in F in yellow. The sire of the scaie bar is 10 Pm.
-
Fig. 15 Neuromuscular preparation of the crayfi~sh slow flexor muscles. A and B reveal
synapsin and dynamin irnrnunolabeling, respectively. Small varicose terminals are stained
by synapsin and dynarnin and these staining pattern CO-localize with one another as shown
in C (regions of CO-localization are shown in yellow). While synapsin stains al1 small
varicosities @), the frequenin labeling is restricted to a subset of the terminal varicosities
(E), as illustrated for the overlay in (F). The size of the scale bar is 10 p.
-
Figure 16 Mature Drosophila Iarval neuromuscula. junction. The synapsin staining is
shown in A, the frequenin staining is shown in B and the overlay of both labels in C in
yellow. Both Ib (marked by the arrow) and 1s (marked by the arrowhead) terminais are
stained. In D the synapsin staining and in E the dynamin staining is revealed. The synapsin
and dynamin staining aimost prefectly ovedap as shown in F. The size of the scaie bar is 20
l u n e
-
Discussion :
a.) Cloning of crayfiih frequenin :
1.) Generai molecular features :
1 have cloned two foms of frequenin, one from lobster and one from crayfish. and
identified them as closely-related members of the family of neuronal calcium sensors (FigA
). The crustacean frequenin molecules are closely related to frequenin from Drosophila
(Fig. 10 and 1 1). The highest degree of conservation at the protein level is found in the EF-
hands, whereas the N- and C-terminal regions are less well conserved. This growing farnily
of neuronal calcium sensor proteins includes the neurocalcins (Dyers et al.. 1996),
recoverin (see Ames et al., 1996, for review) and the frequenins (Pongs et al. 1993,
Olafsson et al., 1995 and 1997). The C-terminus of lobster frequenin is remarkably similar
to that of Drosophila frequenin (Fig. I l) while the N-terminus is not. Interestingly. the
reverse pattern of sequence conservation is found for crafish frequenin (Fig. 10). For
crayfkh frequenin. the N-terminus is more similar to the comparable region of Drosophila
frequenin, while the C-terminus beyond the fourth EF-hand is less well conserved.
The genomic structure of Drosophila frequenin has been described (Pongs et al.,
1993), but the genomic structure of other frequenins remains to be elucidated. This is of
great interest from an evolutionary standpoint. With further observations, it will be possible
-
to determine if the frequenins arose from a common ancestor and the variable N- and C-
terminus of the different members was added by exon shuffling to the core exon encoding
the EF-hands (Gilbert et ai., 1997).
II.) Conservation of EF-bands :
The conservation of the EF-hands underlines their importance for calcium-binding of
these proteins. In crayfikh and lobster frequenin as well as in other frequenins descnbed so
far (Pongs et al., 1993 and (Olafsson et al., 1995 and 1997), the first and fourth EF-han&
are disabled. The second and third EF-hands, however, are remarkably well conserved
except for a single arnino-acid change (Y+E=). Pongs et al. (1993) confirmed the
importance of the second and third EF-han& of Drosophiia frequenin for ca2+-binding by
systematically mutating the EF-hands and perforrning ca2+-induced rnobility shift assays to
assess these mutants for their ability to bind ca2+. They found that mutation of the second
and third EF-hands of Drusophila frequenin criticdly affected the ability of the protein to
bind ~ a ?
III.) N-myristoylation motif :
Crayfi~sh frequenin has the classical N-terminal myristoylation motif (MGSSX) (Ames et
al., 1996) at the N-terminus (Fig. 8), while Iobster frequenin lacks the motif (Fig. 9).
-
Whether a glycine residue further downstream at the N-terminus of lobster frequenin is
myristoylated remains to be shown.
For recovenn, the N-myristoylation is of functional importance as a conformational
switch (Ames et al., 1996). Upon calcium-binding, the myristoyl group of recovenn
becornes exposed on the protein backbone and dlows the protein now to interact with its
targets and the membrane. So far, Drosophilu (Pongs et al., 1993), Xenopus (Olafsson et
al., 1995), mouse (Olafsson et al., 1997) and crayfkh frequenin molecules (this study) carry
the classical myristoylation motif. Only lobster frequenin lacks this motif. It will be
interesting to determine the function of myristoylation of frequenin in tems of its ability to
bind caZC and to bind to membranes.
IV.) Functional implications derived frorn the prhary sequence :
The importance of this degree of sequence flexibility of the N- and C-terminus is
currently not understood. Even though the pnmary structure of the N- and C-terminus is
different arnong the family members, it might be that they d l share common eiements of
secondary structure. It will be very interesting to study the structural dynamics of frequenin
upon calcium binding and its calcium binding characteristics. The primary structure of
lobster and crayfish frequenin predicts two functional ca2+-binding sites, as the first and
-
fourth EF-han& are predicted to be non-hinctionai based on the primary arnino acid
sequence.
These studies will be of importance not only from a stmcturd point of view, but also to
better understand the physiological function of the protein and its possible regdation of the
intraterminal calcium dynamics.
Modeling s ~ d i e s predict that the secondary structure of crayfiish frequenin is similar to
that of recoverin (Fîg. 17) other membea of the family of calcium-binding proteins (Arnes
et ai., 1996) and contains 8 helices, 2 strands and 12 p-tums. As the C-terminai sequence
3' to the fourth EF-hand is however distinct from recoverin, this sequence was not included
in the modeling of the crayfish frequenin structure. To confirm this predicted mode1 and to
resolve further structural details, structural spectroscopic snidies of crayf~sh frequenin are
required.
V.) Future work :
The cloning of crayfish frequenin will also allow us to further address and understand
its precise physiological function. Injection of the recombinant protein into phasic and
tonic motor neurons will help us to understand the function of frequenin in transmitter
release. For this type of expenment, non- and myristoylated recombinant frequenin can be
produced in E. coli and injected into the axons of 'phasic' and 'tonic* motor neurons close
-
to the transrnitter release sites. It would be predicted that the injection of frequenin into
axons of tonic motor neurons wilI enhance neurotransmitter release.
The generation of antibodies against crayfish Frequenin will not only be instrumental
for future comparative labeling experiments, but also complement the protein injection
experiments. Injection of whole anticrayfi~sh igG molecules or Fab' fragments rnight
hinctionally deplete frequenin from crayfish terminais. We will then be able to study the
consequences of this depletion on transrnitter release in the two different types of
motoneuron terminds. If frequenin enhances synaptic eficacy, its functionai depletion
from synaptic terminals would be predicted to decrease the probability of transrnitter
release.
Such antibody inhibition experiments have been successfilly carried out to probe
the rote of spapsin (Pieribone et al. 1996) in neurotransrnitter release. These injection
experiments of fluorophoreconjugated anti-synapsin antibodies also allowed the
identification of two functionally distinct pools of synaptic vesicles and their movement to
be followed in a live preparation. The importance of dynamin in neurotransrnitter release
has been most recently studied by disrupting its interaction with one of its binding partners,
amphiphysin (Shupliakov et ai. 1997). Amphiphysin, which occurs in various isofoms,
binds dynamin via its SH3 domain near its COOH terminus. Microinjection of a
-
fluorophore-conjugated fusion protein containing the SH3 domain of human arnphiphysin
inhibited synaptic vesicle endocytosis which resulted in an activity-dependent distortion of
the synaptic architecture and a depression of transmitter release.
b. ) Immunolocalization of frequenin at the NMJ :
1.) Patterns of locahation :
We have used the cross-reactivity of the anti-Drosophiln frequenin antibody as an
indication that frequenin-like molecules exist at crayfiish neuromuscular junctions. This
does not provide final proof of the existence of frequenin at these synapses, but implies that
closely-related molecules are expressed. To further strengthen the notion that crayfish
frequenin is expressed at crayfish neuromuscular junctions, the anti-Drosophila frequenin
antibody has to be tested for cross-reactivity with crayfïsh frequenin cloned in this study
and its staining pattern compared to that revealed with the anti-crayfish frequenin antibody.
We have shown that frequenin is heavily expressed in thin filiform and less strongly in
more varicose motor neuron terminals (Figs. 13, 14 and 15). These thin filiform terminals
could be identified as belonging to the 'phasic' axon, while the more varicose type of
terminals belong to the 'tonic' axon. The thin filiforni terminals have been shown to
-
release much more neurotransmitter per impulse than the thicker varicose terminals, and to
generate the large 'phasic' EPSPs (Bradacs et al., 1997).
In the almost exclusively 'phasically' innervated crayfish abdominal fast extensor
muscle (Parnas and Atwood, 1966)- frequenin was found to be expressed in a dense
network of 'phasic' terminais, which completely overlapped with the expression pattern of
synapsin imrnunoreactivity (Fig. 14 b.) and c.)). In the dually innervated leg extensor
muscle (Atwood and Cooper, 1995), frequenin was found to be heavily expressed in the
'phasic' terminals, but less strongly in the 'tonic' terminals (Fig. 13 b.) and c.)). In the slow
abdominal flexor muscle preparation, frequenin was found to be expressed in a subset of
the more varicose terminals of the 'tonic' type. These terminals may belong to axon 6.
Although this axon is of the 'tonic' type, it has a more intermittent pattern of impulse
production and is less 'tonic' in its activity than the other motor neurons supplying this
muscle m i n e et ai., 1974).
Expression of synaptic vesicle proteins such as synapsin and dynamin are detected in
both 'phasic' and 'tonic' nerve terminals, suggesting that there are not gross detectable
differences (at the light rnicroscopic level) between 'phasic' and 'tonic' motor neurons in
these proteins. At the mature Drosophi!~ larval neuromuscluar junction (NMJ), frequenin
was expressed in both Ib and Is terminais. These terrninals are glutamatergic and differ in
-
their transmitter release properties albeit the differences are less pronounced than for the
'phasic' and 'tonic' crustacean motoneurons (Atwood et al. ,1993 and Atwood and Cooper,
1995). At the Drosophila NMJ, synapsin and dynamin were also both expressed in Ib and
1s boutons. The localization of synapsin and dynamin at the fly larval neuromuscular
junction almost entirely overlapped with the expression pattern of frequenin, which is also
expressed in Ib and 1s boutons.
II.) Functional implications : physiology
As mentioned above, frequenin was heavily expressed in 'phasic' terminals and less
strongly in 'tonic' terminals in the leg extensor muscle preparation. Freq